2 citations · 11 across the 15 of their papers we have counts for
15 papers
Self-Distillation for Model Stacking Unlocks Cross-Lingual NLU in 200+ Languages
Fabian David Schmidt, Philipp Borchert, Ivan Vulić +1
LLMs have become a go-to solution not just for text generation, but also for natural language understanding (NLU) tasks. Acquiring extensive knowledge through language modeling on…
TopViewRS: Vision-Language Models as Top-View Spatial Reasoners
Chengzu Li, Caiqi Zhang, Han Zhou +3
Top-view perspective denotes a typical way in which humans read and reason over different types of maps, and it is vital for localization and navigation of humans as well as of `no…
Segment Any Text: A Universal Approach for Robust, Efficient and Adaptable Sentence Segmentation
Markus Frohmann, Igor Sterner, Ivan Vulić +2
Segmenting text into sentences plays an early and crucial role in many NLP systems. This is commonly achieved by using rule-based or statistical methods relying on lexical features…
Fairer Preferences Elicit Improved Human-Aligned Large Language Model Judgments
Han Zhou, Xingchen Wan, Yinhong Liu +3
Large language models (LLMs) have shown promising abilities as cost-effective and reference-free evaluators for assessing language generation quality. In particular, pairwise LLM e…
Specialising and Analysing Instruction-Tuned and Byte-Level Language Models for Organic Reaction Prediction
Jiayun Pang, Ivan Vulić
Transformer-based encoder-decoder models have demonstrated impressive results in chemical reaction prediction tasks. However, these models typically rely on pretraining using tens…
Zero-Shot Tokenizer Transfer
Benjamin Minixhofer, Edoardo Maria Ponti, Ivan Vulić
Language models (LMs) are bound to their tokenizer, which maps raw text to a sequence of vocabulary items (tokens). This restricts their flexibility: for example, LMs trained prima…
CALRec: Contrastive Alignment of Generative LLMs for Sequential Recommendation
Yaoyiran Li, Xiang Zhai, Moustafa Alzantot +4
Traditional recommender systems such as matrix factorization methods have primarily focused on learning a shared dense embedding space to represent both items and user preferences.…
Analyzing and Adapting Large Language Models for Few-Shot Multilingual NLU: Are We There Yet?
Evgeniia Razumovskaia, Ivan Vulić, Anna Korhonen
Supervised fine-tuning (SFT), supervised instruction tuning (SIT) and in-context learning (ICL) are three alternative, de facto standard approaches to few-shot learning. ICL has ga…
Scaling Sparse Fine-Tuning to Large Language Models
Alan Ansell, Ivan Vulić, Hannah Sterz +2
Large Language Models (LLMs) are difficult to fully fine-tune (e.g., with instructions or human feedback) due to their sheer number of parameters. A family of parameter-efficient s…
UNDIAL: Self-Distillation with Adjusted Logits for Robust Unlearning in Large Language Models
Yijiang River Dong, Hongzhou Lin, Mikhail Belkin +2
Mitigating the retention of sensitive or private information in large language models is essential for enhancing privacy and safety. Existing unlearning methods, like Gradient Asce…
Self-Augmented In-Context Learning for Unsupervised Word Translation
Yaoyiran Li, Anna Korhonen, Ivan Vulić
Recent work has shown that, while large language models (LLMs) demonstrate strong word translation or bilingual lexicon induction (BLI) capabilities in few-shot setups, they still…
Pheme: Efficient and Conversational Speech Generation
Paweł Budzianowski, Taras Sereda, Tomasz Cichy +1
In recent years, speech generation has seen remarkable progress, now achieving one-shot generation capability that is often virtually indistinguishable from real human voice. Integ…
DIALIGHT: Lightweight Multilingual Development and Evaluation of Task-Oriented Dialogue Systems with Large Language Models
Songbo Hu, Xiaobin Wang, Zhangdie Yuan +2
We present DIALIGHT, a toolkit for developing and evaluating multilingual Task-Oriented Dialogue (ToD) systems which facilitates systematic evaluations and comparisons between ToD…
Large Language Models are Miscalibrated In-Context Learners
Chengzu Li, Han Zhou, Goran Glavaš +2
When adapting ICL with or without fine-tuning, we are curious about whether the instruction-tuned language model is able to achieve well-calibrated results without suffering from t…
Adapters: A Unified Library for Parameter-Efficient and Modular Transfer Learning
Clifton Poth, Hannah Sterz, Indraneil Paul +7
We introduce Adapters, an open-source library that unifies parameter-efficient and modular transfer learning in large language models. By integrating 10 diverse adapter methods int…